Speckle Observations of Binary Stars with the WIYN Telescope. III. A Partial Survey of A, F, and G Dwarfs

Transcription

1 Rochester Institute of Technology RIT Scholar Works Articles Speckle Observations of Binary Stars with the WIYN Telescope. III. A Partial Survey of A, F, and G Dwarfs Elliott P. Horch Rochester Institute of Technology Sarah Robinson Rochester Institute of Technology Zoran Ninkov Rochester Institute of Technology William F. van Altena Yale University Reed D. Meyer Yale University See next page for additional authors Follow this and additional works at: Recommended Citation Elliott P. Horch et al 2002 AJ This Article is brought to you for free and open access by RIT Scholar Works. It has been accepted for inclusion in Articles by an authorized administrator of RIT Scholar Works. For more information, please contact ritscholarworks@rit.edu.

2 Authors Elliott P. Horch, Sarah Robinson, Zoran Ninkov, William F. van Altena, Reed D. Meyer, Sean E. Urban, and Brian D. Mason This article is available at RIT Scholar Works:

3 The Astronomical Journal, 124: , 2002 October # The American Astronomical Society. All rights reserved. Printed in U.S.A. SPECKLE OBSERVATIONS OF BINARY STARS WITH THE WIYN TELESCOPE. III. A PARTIAL SURVEY OF A, F, AND G DWARFS 1 Elliott P. Horch, 2,3 Sarah E. Robinson, 2 and Zoran Ninkov Chester F. Carlson Center for Imaging Science, Rochester Institute of Technology, 54 Lomb Memorial Drive, Rochester, NY ; horch@cis.rit.edu, ser9180@cis.rit.edu, ninkov@cis.rit.edu William F. van Altena 2 and Reed D. Meyer 2 Department of Astronomy, Yale University, P.O. Box , New Haven, CT ; vanalten@astro.yale.edu, rdm@astro.yale.edu and Sean E. Urban and Brian D. Mason US Naval Observatory, 3450 Massachusetts Avenue, NW, Washington, DC 20392; seu@pyxis.usno.navy.mil, bdm@draco.usno.navy.mil Received 2002 May 13; accepted 2002 June 25 ABSTRACT Two hundred thirty nearby main-sequence stars with spectral types in the range of A to G have been observed by way of speckle interferometry using the WIYN 3.5 m telescope at Kitt Peak, Arizona. The stars had no previous mention of duplicity in the literature. Of those observed, 14 showed clear evidence of a companion, and 63 were classified as suspected nonsingle based on a power spectrum analysis. The remaining stars discussed show no evidence of duplicity to the limit of the detection system in high-quality observations. Key words: astrometry binaries: visual techniques: interferometric 1. INTRODUCTION Although a speckle survey has been completed on a large sample of bright (5.0 < V < 6.5) stars (McAlister et al. 1987, 1989, 1993), field stars in the magnitude range have not been uniformly surveyed for duplicity yet are prime targets for speckle interferometry. Several surveys of specialized groups of stars have been completed, however, such as high-velocity stars (Lu et al. 1987), open cluster stars (Mason et al. 1993a, 1993b; Patience et al. 1998, 2002), Be stars (Mason et al. 1997), star-forming regions (Ghez, Neugebauer, & Matthews 1993; Leinert et al. 1993), white dwarfs (McAlister et al. 1996), O stars (Mason et al. 1998), Wolf-Rayet stars (Hartkopf et al. 1999), and Bootis stars (Marchetti, Faraggiana, & Bonifacio 2001). Together, these surveys have provided important information on binary statistics, as well as binary formation and evolution issues. With the publication of the final Hipparcos Catalogue (ESA 1997), improved distances now exist to nearby stars. If a set of solar-type Hipparcos stars were to be identified as true binaries, they would probably have reasonably short periods, allowing for precise orbit determination within the next couple of decades. This information could then be used to refine the empirical mass-luminosity relation (MLR), for example, which is in need of improvement, especially in the mass range of M (Henry & McCarthy 1993; Henry et al. 1999). This is due to the fact that many stars contributing to this region of the MLR are visual and speckle binaries, whose mass estimates are dominated by parallax uncertainty, even post-hipparcos. However, in view of further improvements in the distances to these systems that are likely from missions such as the Space Interferometry Mission (SIM) and the Global Astrometric Interferometer for Astrophysics, the promise for substantial improvements in the MLR in the next two decades is high. Nearby solar-type stars were considered for the makeup of the astrometric grid for the SIM, although it now appears that most grid stars will be giants that are much farther away (nominally 1 kpc). Nonetheless, duplicity surveys of nearby stars have some utility with regard to SIM in yielding information of the binary statistics of the types of stars that are the progenitors of the giants like those to be used in the grid, as well as providing a sample of secondary candidate grid stars, which could contribute to the performance of the satellite pointing at some level. Another potential use of the objects unresolved in our study would be as science targets for planet searches with SIM or future NASA missions such as the planned Terrestrial Planet Finder. Horch et al. (1999, hereafter Paper I; 2002, hereafter Paper II) demonstrated the basic capabilities of a fast readout CCD camera used as a speckle imaging device at the WIYN telescope. In this paper, we report results of speckle observations of a sample of stars in the distance range of pc and in the spectral range A to G, none of which had previous indications of close (i.e., subarcsecond) companions in the literature. 1 The WIYN Observatory is a joint facility of the University of Wisconsin Madison, Indiana University, Yale University, and the National Optical Astronomy Observatory. 2 Visiting Astronomer, Kitt Peak National Observatory, National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation. 3 Current address: Department of Physics, University of Massachusetts Dartmouth, 285 Old Westport Road, North Dartmouth, MA OBSERVATIONS AND DATA REDUCTION Observations were carried out at the WIYN 3.5 m telescope at Kitt Peak, Arizona. As discussed in Papers I and II, the speckle camera consists of a front-illuminated Photometrics CCD camera with a Kodak KAF-4200 chip and a speckle optics package on loan from J. G. Timothy of Nightsen, Inc. The readout electronics for this system operates at a rate of 500 kpixel s 1, and the method of obtaining small-format ( pixels) speckle frames

4 2246 HORCH ET AL. with the device was the strip method of Horch, Ninkov, & Slawson (1997). A typical observation for the survey described here was a sequence of 700 frames, stored as a stack of approximately 70 strips, with each strip containing eight to 12 images. The strips were then further separated into pixel images, each containing one speckle pattern. The observations were obtained during several runs spanning the time frame 1998 December to 2000 October, and the large majority were obtained in subarcsecond seeing conditions. Analysis of the individual speckle images was carried out as described in Paper I. The technique is based on a study of the spatial frequency power spectrum for each object deconvolved with that of a point source observed close in time and near to the target on the sky. Such calibration point sources are selected from the Bright Star Catalogue (Hoffleit & Jaschek 1982). In the case of a binary star, the deconvolved power spectrum will exhibit a fringe pattern where the spacing and orientation of the fringes is related to the separation and position angle of the two stars on the image plane. In the case of an unresolved star, no such pattern is found, and the deconvolved power spectrum should, in principle, be flat out to the diffraction limit of the power spectrum. Rudimentary reconstructed images were also made using two near-axis subplanes of the image bispectrum and studied for all objects discussed here. These images were useful in confirming the power spectrum result and resolving the quadrant ambiguity inherent in the power-spectrum approach for discovered components. As stated earlier, the target list is a subset of potential SIM astrometric grid candidates. At the time of selection, it was unclear if SIM would use faint, distant grid stars or nearby, bright ones. The bright grid stars would allow a variety of techniques to be used to uncover duplicity, including speckle interferometry, long baseline optical interferometry, CMOS observations (see, e.g., Winter 2000), Fourier transform spectroscopy (FTS), and the utilization of historical observations (Mason et al. 1999). Each works best on brighter objects; this drove the magnitude selection criterion to 6.5 < V < 8.5, where the lower bound (bright end) was set by the SIM program and the upper bound would allow enough stars for a uniform grid of about 7000 objects. One technique (the FTS) works best if there are many spectral lines, so early-type stars were avoided. The final selection criteria were A5 K5 class V with variability less than 0.1 mag and no historical data indicating duplicity. A spatially uniform list was made with priority given to those stars already having at least one large aperture speckle or optical interferometric observation. The WIYN list originally consisted of stars of any allowable spectral classification with a magnitude fainter than However, as the SIM project evolved to place a higher emphasis on G dwarfs, the magnitude restriction was dropped but a G dwarf spectral type restriction was put in place. The final WIYN target list was composed of main-sequence stars with spectral types in the range of A5 G2 and within approximately 250 pc from the Sun. The very great majority of the stars observed fit these criteria. A more detailed explanation of the construction of the full target list and an overview of the observing programs was discussed by Urban et al. (1999). For the part of the project discussed here, a total of over 500 observations were divided into four basic groups in the analysis phase: (1) single to the detection limit (unresolved), if the object showed no evidence of duplicity in a highquality observation as defined below; (2) suspected nonsingle, if the object showed evidence of power spectrum irregularities, that is, features potentially consistent with duplicity but inconclusive in nature; (3) double, if the object showed clear evidence of binary fringes; and (4) no result claimed, if the observation was not high enough in quality to make a definitive statement to the detection limit. We desired the results reported here to be of the highest possible quality, so a signal-to-noise ratio of 3 at half the Nyquist frequency was required for inclusion in the unresolved category. This limit is intentionally conservative so as to provide a list of stars with the highest chance of being single and limits the observations reported here to approximately half of the total taken. Of course, even this criterion does not guarantee that a star is single, because detection capabilities impose limitations on the magnitude difference and total magnitude that can be successfully measured as discussed below, and because it is possible for the orbital geometry to be such that even an easily resolved system may be below the diffraction limit during the epoch of observation. 3. RESULTS Of the stars observed, 153 were high enough in quality to be classified as unresolved, and 63 were classified as suspected nonsingle. Of these, five objects observed twice had one observation resulting in a determination of unresolved and one resulting in a determination of suspected nonsingle. Fourteen stars were classified as double. The remainder fell into category 4 described above and therefore are not reported. The results for the other three categories are displayed in Tables 1 3. Table 1 shows observations of stars in category 1 (unresolved), for which the column headings are (1) the Durchmusterung number (BD in all cases here); (2) the HD number; (3) the Hipparcos or Tycho Catalogue number; (4) the right ascension and declination in J coordinates; (5) the V magnitude of the target; (6) the observation date in fraction of the Besselian year; (7) the center wavelength of the filter used to make the observation, in nanometers; and (8) the full width at half maximum of the filter passband, also in nanometers. Table 2 shows the suspected nonsingle stars, for which the column headings are identical to Table 1. Five objects observed twice have one entry in Table 1 and the other in Table 2. While a lower quality observation could produce a spurious suspected nonsingle result, it is also possible that orbital motion of a companion with separation below the diffraction limit might cause a suspected nonsingle result during the portion of the orbit with relatively large separation and a single result during the portion of relatively small separation. We suggest that the overall conclusion for these objects at the moment must therefore be that they are suspected nonsingle pending future observations. Table 3 shows the 14 objects resolved as doubles in our survey; the discoverer designation of Yale-RIT (YR) 19 through 32 is given here. In this case, the column headings are (1) the Durchmusterung number (BD in all cases); (2) the discoverer designation; (3) the HD number; (4) the Hipparcos or Tycho Catalogue number; (5) the right ascension and declination in J coordinates, (6) the system V magnitude; (7) the observation date in fraction of the Besselian year; (8) the observed position angle (h), in degrees, with north through east defining the positive sense of h; (9) the observed separation () in arcseconds; (10) the

5 TABLE 1 Stars Showing No Evidence of Duplicity in High-Quality Observations DM (1) HD (2) HIP or TYC (3), (J2000.0) (4) V (5) Date (6) (7) D (8) BD BD BD BD BD BD BD BD a BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD a BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD

6 TABLE 1 Continued DM (1) HD (2) HIP or TYC (3), (J2000.0) (4) V (5) Date (6) (7) D (8) BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD

7 SPECKLE OBSERVATIONS OF BINARY STARS. III TABLE 1 Continued DM (1) HD (2) HIP or TYC (3), (J2000.0) (4) V (5) Date (6) (7) D (8) BD BD BD a BD BD BD BD BD BD a BD BD BD BD BD BD BD BD a BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD a Another observation of this object is reported in Table 2. center wavelength of the filter used to make the observation, in nanometers; and (11) the full width at half-maximum of the filter passband, also in nanometers. The position angles have not been corrected for precession and are appropriate for the epoch of observation shown. Position angles and separations are shown without uncertainties, but a reasonable uncertainty estimate for any measure shown may be obtained from the discussion in Paper II on measurement precision. In general, position angle measurements with our system are made with a precision of approximately 1 and separation measures with a precision of 3 mas, although these figures will vary depending on magnitude, separation, and other factors, as explained in Paper II. Because of the large format of the CCD used and the details of the strip method of acquiring speckle data, there is a small possibility that a large separation component that lands on the chip but outside the strip could be aliased to a subarcsecond separation. However, no such companions were noted while examining short full-frame exposures of the targets during the observations, and the closest companions to any of the systems in Table 3 listed in the SIMBAD database are approximately 2 0 away from the targets here. These wide components would therefore be safely off the CCD active area. The speckle instrument has a field of view of approximately ; half of the active area is masked off for speckle observations. Although the measures in Table 3 do not include estimates of the magnitude difference of each system, it is possible to derive these from the power spectrum analysis in addition to the astrometry. For the systems described here, the raw magnitude differences spanned a range from 0.0 to 3.7 mag. Because of the opportunity to better characterize the photometric precision with a large sample of observations, magnitude differences for these 14 systems will be presented in a future publication devoted to speckle photometry at WIYN (Horch, Meyer, & van Altena 2002). Observations that led to an ambiguous quadrant determination for the secondary are noted in Table 3. Generally, these

8 TABLE 2 Suspected Nonsingle Stars DM (1) HD (2) HIP or TYC (3), (J2000.0) (4) V (5) Date (6) (7) D (8) BD BD a BD BD BD BD BD BD BD BD BD BD a BD BD BD BD BD BD * BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD BD a BD BD BD a BD BD BD a BD BD BD BD a Another observation of this object is reported in Table 1.

9 SPECKLE OBSERVATIONS OF BINARY STARS. III TABLE 3 Speckle Relative Astrometry Measures for Newly Resolved Pairs DM Discoverer Designation HD HIP or TYC, (J2000.0) V Date h (deg) (arcsec) D BD YR a BD YR BD YR b BD YR BD YR BD YR BD YR BD YR BD YR BD YR a BD YR BD YR a BD YR BD YR a Quadrant ambiguous. b Quadrant ambiguous and evidently inconsistent with the previous observation. objects have smaller separations, smaller magnitude differences, or both. Signal-to-noise ratio is another factor that can lead to a lack of a definitive quadrant determination from bispectral analysis. Figure 1 shows two examples of our rudimentary reconstructed images obtained for resolved doubles. It should be emphasized that these images are presented as an illustration of data quality only and not as meaningful irradiance maps. Only two subplanes of the image bispectrum were used in the construction of phase map in the Fourier plane, and we started with the zero phase assumption and iterated from that point. This has the tendency of often leaving a ghost peak in the reconstructed image in the opposite quadrant from the true secondary location. In most cases, the amplitude of this peak is low enough that no confusion exists in the determination of the true secondary location, but it renders the photometric information suspect. The resolution of the images is also typically not diffraction-limited; we filter the reconstructed Fourier transform of the object with a low-pass Gaussian filter in order to attenuate noise. These features are in contrast to the reconstructed images that were presented in Paper II; in that work, we used many Fig. 1. Examples of rudimentary reconstructed images made in the process of the data analysis for two resolved double stars: (a) YR 29 (=HD ) and (b) YR 32 (=HD ). In both plots, solid contours are drawn at 0.02, 0.04, 0.06, 0.1, 0.3, and 0.5 of the maximum value in the array, and dotted contours are drawn at 0.02 and A low ghost peak in the opposite quadrant from the secondary is present in both images because of the zero-phase input assumption of the reconstruction technique. The YR 29 image is one of the higher quality images obtained from the 14 resolved systems, and the YR 31 image is closer to average quality.

10 2252 HORCH ET AL. Vol. 124 Fig. 2. Binary detection capabilities of the speckle system at the WIYN telescope in terms of magnitude difference and separation of the companion. In both plots, the open circles are known binaries observed at WIYN over the same time period as the survey objects and successfully reduced. (Astrometry for these systems appears in Paper II, and in nearly all cases the Dm coordinate shown here is that appearing in the Hipparcos Catalogue.) The filled circles are the locations of the 14 newly discovered doubles in Table 3 (where the separation coordinate is that appearing in Table 3, and the magnitude difference is the preliminary result from the speckle analysis). (a) Objects in the V magnitude range (b) Objects in the V magnitude range more subplanes in order to make a more robust phase estimate and did not use the zero-phase starting point. Paper II also discussed detection limits of the speckle system at WIYN (particularly Fig. 7 and its description), finding that systems with magnitude differences of up to four and system magnitudes of as faint as 10 can normally be detected and reduced successfully in high-quality observations. The sample of target stars in this survey have total magnitudes in the range , well within the system magnitude limit. In Figure 2, we show the separation and magnitude difference coordinates for all known binaries resolved at WIYN over the same time frame as the survey project described here. Two magnitude bins are shown; in Figure 2a, stars with system magnitudes between 6.5 and 7.5 are shown, while in Figure 2b, stars with system magnitudes between 7.5 and 8.5 are shown. Using column (5) in Tables 1 and 2, one can easily estimate the detection limits for any object presented here. These plots no doubt contain a bias in the WIYN binary target list toward smaller magnitude differences at small separation, since there are fewer large Dm, small separation systems listed in the Washington Double Star Catalog. Nonetheless, the envelope of these points represents a conservative lower limit of detectability for companions for each magnitude bin. All stars in Table 1, if binary or multiple, would therefore have an expected separation and/or magnitude difference outside of the envelope. Similarly, any object in Table 2 that is truly binary or multiple would have an expected position near the edge of the envelope. The locations of the 16 observations listed in Table 3 are also plotted in the figures. As their locations indicate, these are detections within the limits expected for the system, with most systems near the envelope boundary. 4. DISCUSSION These results can be compared with the earlier survey of McAlister et al. (1987, 1989, 1993), where those authors observed over 2000 objects from the Bright Star Catalogue in an attempt to determine the duplicity fraction, as well as new candidate systems for astrophysical studies. In that work, a binary fraction of approximately 15% was seen for dwarfs regardless of spectral type, including binaries known from previous observing techniques. In the first of the three mentioned references, 39 of 426 dwarfs (i.e., luminosity class V only) were resolved for the first time via speckle, but in the third reference, some of these discoveries were then determined to be spurious. Removing these from the list, 26 dwarf doubles remain, and therefore the fraction of stars determined to be double from speckle observations was approximately 6.1%. Using the figures in our survey, we obtain the same speckle discovery rate of 14/230 =6.1%. Due to the selection criteria discussed in x 2, we may view the current survey as a collection of similar, albeit fainter, stars. While the range of distances was larger in the McAlister et al. sample, the average is comparable to that of the systems here, and the difference in brightness is largely due to a shift toward later spectral types in our survey. Since McAlister et al. found little variation in the binary fraction with spectral type for dwarfs, we therefore expect a similar binary fraction if the detection efficiencies are similar. The speckle camera used by those authors at the time had a limiting magnitude of approximately 8.5 at the Kitt Peak 4 m telescope, whereas the system used here has a limiting magnitude of at least 10 at WIYN and is sensitive to somewhat larger magnitude differences. In both studies, the objects in the survey were therefore 2 3 mag brighter than the detection limit, and any objects resolved because of the greater sensitivity to larger magnitude differences at WIYN may well be offset by the larger diffraction limit of the slightly smaller telescope aperture and redder filter used for many of our observations. Overall, we conclude that a similarity of discovery rate is plausible from instrumental, as well as astrophysical, considerations. If the detection efficiency is, in fact, comparable between the two studies, it may also suggest that few of the suspected nonsingle stars reported here are actually resolvable at WIYN, although more observations are needed to confirm this hypothesis.

11 No. 4, 2002 SPECKLE OBSERVATIONS OF BINARY STARS. III CONCLUSIONS We have reported the results of speckle observations of 230 field stars with spectral types ranging from A to G examined using CCD-based speckle interferometry at the WIYN 3.5 m telescope. Of these stars, 153 were found to have no evidence of duplicity in high-quality observations, 63 showed signatures in the power spectrum that indicated a difference from a single star power spectrum and are reported as suspected nonsingle, and 14 stars were resolved for the first time and found to be double. The remaining stars observed in the survey did not meet the quality criteria for inclusion in the group of high-quality nondetections. The discovery rate is consistent with an earlier survey on bright stars by McAlister et al. We are grateful to J. Gethyn Timothy for continued use of the speckle optics package and to Charles Corson, David Sawyer, Wendy Shook, Gillian Rosenstein, and Eugene McDougall at Kitt Peak for their excellent assistance with the telescope, as well as the referee. This work was funded by JPL subcontract from the Preparatory Science Program for the Space Interferometry Mission, and a Theodore Dunham, Jr. Grant from the Fund for Astrophysical Research. It also made use of the Washington Double Star Catalog, maintained at the US Naval Observatory, and the SIMBAD database, operated at CDS, Strasbourg, France. ESA. 1997, The Hipparcos and Tycho Catalogues (ESA SP-1200) (Noordwijk: ESA) Ghez, A. M., Neugebauer, G., & Matthews, K. 1993, AJ, 106, 2005 Hartkopf, W. I., Mason, B. D., Gies, D. R., ten Brummelaar, T., McAlister, H. A., Moffat, A. F. J., Shara, M. M., & Wallace, D. J. 1999, AJ, 118, 509 Henry, T. J., Franz, O. G., Wasserman, L. H., Benedict, G. F., Shelus, P. J., Ianna, P. A., Kirkpatrick, J. D., & McCarthy, D. W., Jr. 1999, ApJ, 512, 864 Henry, T. J., & McCarthy, D. W., Jr. 1993, AJ, 106, 773 Hoffleit, D., & Jaschek, C. 1982, The Bright Star Catalogue (4th rev. ed.; New Haven: Yale Univ. Obs.) Horch, E., Ninkov, Z., van Altena, W. F., Meyer, R. D., Girard, T. M., & Timothy, J. G. 1999, AJ, 117, 548 (Paper I) Horch, E. P., Meyer, R. D., & van Altena, W. F. 2002, in preparation Horch, E. P., Ninkov, Z., & Slawson, R. W. 1997, AJ, 114, 2117 Horch, E. P., Robinson, S. E., Meyer, R. D., van Altena, W. F., Ninkov, Z., & Piterman, A. 2002, AJ, 123, 3442 (Paper II) Leinert, Ch., Zinnecker, H., Weitzel, N., Christou, J., Ridgway, S. T., Jameson, R., Haas, M., & Lenzen, R. 1993, A&A, 278, 129 Lu, P. K., Demarque, P., van Altena, W., McAlister, H., & Hartkopf, W. 1987, AJ, 94, 1318 Marchetti, E., Faraggiana, R., & Bonifacio, P. 2001, A&A, 370, 524 Mason, B. D., Corbin, T. E., Hajian, A. R., Hummel, C. A., Rafferty, T. J., Urban, S. E., & Zacharias, N. 1999, BAAS, 31, REFERENCES Mason, B. D., Gies, D. R., Hartkopf, W. I., Bagnuolo, W. G., Jr., ten Brummelaar, T., & McAlister, H. A. 1998, AJ, 115, 821 Mason, B. D., Hartkopf, W. I., McAlister, H. A., & Sowell, J. R. 1993a, AJ, 106, 637 Mason, B. D., McAlister, H. A., Hartkopf, W. I., & Bagnuolo, W. G., Jr. 1993b, AJ, 105, 220 Mason, B. D., ten Brummelaar, T., Gies, D. R., Hartkopf, W. I., & Thaller, M. L. 1997, AJ, 114, 2112 McAlister, H. A., Hartkopf, W. I., Hutter, D. J., Shara, M. M., & Franz, O. G. 1987, AJ, 93, 183 McAlister, H. A., Hartkopf, W. I., Sowell, J. R., Dombrowski, E. G., & Franz, O. G. 1989, AJ, 97, 510 McAlister, H. A., Mason, B. D., Hartkopf, W. I., Roberts, L. C., Jr., & Shara, M. M. 1996, AJ, 112, 1169 McAlister, H. A., Mason, B. D., Hartkopf, W. I., & Shara, M. M. 1993, AJ, 106, 1639 Patience, J., Ghez, A. M., Reid, I. N., & Matthews, K. 2002, AJ, 123, 1570 Patience, J., Ghez, A. M., Reid, I. N., Weinberger, A. J., & Matthews, K. 1998, AJ, 115, 1972 Urban, S., Mason, B., Horch, E., Holdenried, E., Rafferty, T., Hartkopf, W., & van Altena, W. 1999, BAAS, 31, 1440 Winter, L. 2000, in IAU Colloq. 180, Towards Models and Constants for Sub-Microarcsecond Astrometry, ed. K. J. Johnston, D. D. McCarthy, B. J. Luzum, & G. H. Kaplan (Washington: US Naval Obs.), 380